Sample Size Determination
Failure rates of 42% and 27% were expected in the placebo and the
fibrinogen groups, respectively, i.e. corresponding to a 15% absolute
difference 19. Assuming a 5% drop-out rate due
difficulties to collect data in emergency conditions, 434 patients
needed to be enrolled to achieve a power of 90% at the 2-sided
significance level of 0.05. As there was uncertainty about expected
rates which impacted the variance of the failure rate difference, a
blinded interim analysis to reassess sample size to maintain a power of
90% led to a sample size increase from 434 to 470 patients. Given the
lower than expected drop-out rate of 1.55%, the trial was stopped on
June 30, 2018 with 448 patients enrolled and a trial power of 89.6%.